Infrared Small Target Detection Based on Double Gray-Values Descend Angle Contrast Measure

Infrared small target detection remains a subject of significant theoretical research value and practical application potential. However, the complexity of the detection environment and the interference from salt noise considerably increase the difficulty of the detection task. In scenarios involving small infrared target detection against complex backgrounds, existing methods face substantial challenges in achieving an optimal balance between superior detection capability and real-time performance. Furthermore, the influence of noise on detection efficacy is an unavoidable issue. To address these challenges, this paper proposes an anti-noise detection method that excels in both real-time performance and detection ability. Initially, energy residuals are computed to pre-assess potential target areas. Subsequently, the minimum difference between the central pixel and its local neighborhood is calculated to position the target region positioning twice, thereby enhancing targets while suppressing background noise. Finally, by defining a double gray-values descend angle and calculating its tangent quotient, further enhancement of targets is achieved alongside suppression of additional background interference and salt noise. Experimental evaluations were conducted on four publicly available datasets, comparing our proposed approach with seven existing algorithms. The results demonstrate that our method facilitates rapid detection of small targets affected by salt noise within complex backgrounds while exhibiting exceptional detection capabilities and real-time performance. Notably, DGDACM achieves its fastest single-frame detection time at just 0.0719 seconds—5.1% faster than the second-best average single-frame time.